Font Size: a A A

Research On Target Selection Technology Based On Voice Interaction

Posted on:2020-08-23Degree:MasterType:Thesis
Country:ChinaCandidate:H L XieFull Text:PDF
GTID:2438330599955738Subject:Computer system architecture
Abstract/Summary:PDF Full Text Request
Mouse is the mainstream interactive device in the modern human-computer interaction interface,which is used for target selection and positioning,but people with diseases or disabilities cannot use it normally,and normal users may encounter difficulties in use under certain circumstances.At the same time,with the development of computer interaction technology towards nature and intelligence,the audience of computers is more and more extensive.Therefore,it is necessary to study new alternatives to target selection and positioning,so that people benefit from the development of computer interaction technologyAccording to the survey,a variety of solutions have been proposed,among which the voice cursor control scheme is the most practical.Existing voice cursor controls can be divided into two types:one based on direction and the other based on goals.For a direction-based solution,the operation is too slow and is prone to fatigue for the user.For target-based scenarios,if there are duplicate or unnamed targets,the scenario will not be selected.Grid-based voice cursor control is a technology that uses grids for target positioning,which can effectively combine the advantages of the first two schemes,but the target scheme of this type of basic scheme is inefficient,making it difficult to use normally.Aiming at the above problems,this thesis studies a technique of using voice commands to combined with the size of the screen area to select the target,and a technique for using the voice direction command to perform target search and selection.The purpose of the research is to reduce the target selection and positioning task operation steps and reduce the number of voice command releases,improve the efficiency and accuracy of target selection,and user satisfaction.The main work is as follows(1)Two speech selection techniques are proposed:grid-based region selection and target search technology.In order to achieve the above research purposes,two new types of speech cursors are designed:area cursor and ray cursor.The area selection technique based on the area cursor expands the user's selectable area,enabling the user to intuitively select the target directly through the grid area.Based on the target search technology of the ray cursor,a voice direction command is added to enable the user to search for the target position through the voice direction command,and quickly move the cursor to the target in any ray direction.These two technologies are more intuitive and natural,reducing task steps,reducing the number of voice commands,improving the efficiency and accuracy of target selection,and making users more satisfied.(2)In order to verify that the scheme based on Region Selection and Target Search is more efficient than the basic scheme,a comparative experiment is designed to compare the efficiency of the scheme.The experiment analyzed the target selection time,accuracy and user satisfaction,and designed the questionnaire to consider the ease of use and efficiency of the program.The experimental data was then analyzed.The experimental results show that the Region Selection scheme is nearly 38%faster than the basic scheme,and the Target Search is nearly 56%faster than the basic scheme.The accuracy of the three schemes was 81.39%,95.83%,and 97.78%,respectively.Finally,the proposed solution consumes less time,accuracy and user satisfaction when performing target selection tasks.(3)In order to verify the applicability of the target selection technique,a program comparison experiment was designed.The experiment focused on two key factors,target size and target density,and the impact on program efficiency.The experimental results show that the basic scheme is most affected by the target size.The Region Selection and Target Search can effectively overcome the impact of the target size,and the technical stability is higher.For the target density change,Target Search performance is relatively stable,and the degree of impact is small.On the contrary,Region Selection and the basic scheme are not effective.Finally,we summarize the target selection of the three schemes under different conditions,and draw conclusions about the applicable conditions and applicability of the relevant technologies.(4)A speech selection technology application strategy is proposed.In order to improve the efficiency and accuracy of the target selection task of the speech cursor control system,the efficiency analysis of the six representative speech selection technologies in the system was carried out,and the main factors affecting each technology were summarized.Further,based on experimental findings,an application strategy of speech selection technology is proposed:for a simple target environment,consider using direction-based and goal-based techniques.For more complex target environments,consider grid-based technologies.By applying the strategy,the target selection and positioning task operation steps can be effectively reduced and the number of times the voice command is released can be reduced,and the target selection efficiency and accuracy can be improved.To this end,it is recommended to use the application strategy as a design guide to guide the overall design of the voice cursor control system.
Keywords/Search Tags:human-computer interaction, speech interaction, speech recognition, voice cursor, target selection
PDF Full Text Request
Related items